X Ads Account Build With a Six-Week Proving Period

X is a build, not a subscription. We construct the account, run a six-week proving period at an agreed spend, and hand you a written verdict on whether X deserves a standing line in the budget. On roughly half the industrial accounts we have taken through this, the verdict was no, and what you bought was the evidence to stop having the argument internally. The proving period, defined before anything launches: We agree the number X has to hit to survive. Cost per qualified enquiry, or contribution margin on settled orders for DTC, measured in your CRM or payment provider rather than in X Ads Manager. Six weeks, a spend level capable of producing a readable result, and a decision date already in the calendar. Nobody gets to move the goalposts in week five, including us. The build: Objective chosen against the event that genuinely exists on your site. Website Conversions only where a defined conversion event is firing and matching; Traffic where it is not; never Engagements standing in for either, because an engagement objective will always deliver engagements and tell you nothing. Conversion tracking through the X Website Tag with server-side events sent to X's Conversions API from the same event spine feeding your other channels, so the platforms can be compared instead of each grading its own exam. Targeting worth the effort on this platform: Keyword targeting against what people write rather than what they are inferred to like, including competitor product names, error messages, tender and procurement language, and the phrases your sales team hears on first calls. Conversation topics as the broader layer above it. Follower look-alikes built from a hand-assembled list of accounts your buyers actually read, which for South African industrial and DTC categories is shorter than anyone expects and takes an afternoon of genuine research rather than a keyword tool. Attribution settings, corrected on day one: X's defaults credit generously. We cut the post-view window to one day, hold clicks to fourteen, and reconcile platform-reported conversions against the CRM weekly through the test. A platform counting its own contribution will always find a healthy number; the point of six weeks is to learn what X contributed when it is not the one counting. Brand safety: Sensitive content settings, adjacency controls and a category-specific keyword blocklist, configured before launch and reviewed again at week three. We will tell you plainly what these controls do and do not cover on this platform. If your brand cannot tolerate an unpredictable adjacency, that is a legitimate position, and the correct decision is not to advertise here. We will say that on the first call rather than after your money is gone. What we refuse: Takeover products. Timeline Takeover and Trend Takeover are reach purchases priced for a national launch, and we have never seen a case for them in this client profile. Bidding into a trending topic for borrowed relevance. Automated reply campaigns. And running X defensively because a competitor is loud there, which is a decision about someone's ego rather than about your pipeline. Not included: Organic account management, replies outside ad units, creative production beyond assembly and copywriting, and handle or profile strategy. Who this suits, and who it does not: It suits products with a live public conversation attached: developer-adjacent tools, consumer brands with an argument, and financial and sport-linked categories in this market. It does not suit a business selling into a procurement function that does not read X. If your buyer is a plant engineer specifying equipment, that person is reachable on LinkedIn, and we would rather build you that than take your money for this.